All Times ET
Program is Subject to Change
Bayesian Methods Applied to Small Area Estimation for Establishment Statistics (308288)*Scott Holan, US Census Bureau/ University of Missouri
Keywords: Dependent data, Dimension reduction, Hierarchical modeling, Mixed model, Small area estimation
Small area statistics is an ongoing topic of interest to both data-users and official statistical agencies. In contrast to house-hold surveys, establishment surveys present several unique challenges. This overview will provide background on Bayesian methods for small area estimation and highlight the utility of hierarchical modeling for estimation of establishment statistics. Within the hierarchical framework we will describe approaches to incorporating various sources of dependence and methods for combining data from multiple surveys and or auxiliary data sources. As a byproduct of this approach more precise estimation is typically achieved along with quantification of uncertainty. Finally, this overview will provide discussion regarding dimension reduction, distribution theory, and computation.